Contents
What are the best practices for de novo transcriptome assembly?
While early genome assemblers used pairwise overlaps between long reads to extend contigs, this approach is unfeasible when dealing with hundreds of millions of reads. Thus, de novo transcriptome asssemblers use DeBruijn graphs, which are constructed and extended based upon kmers, i.e. subsequence of length k found in reads.
Can a lowly expressed transcript be assembly tool?
In theory, lowly expressed transcripts may lead to the occurence of biologically real, rare kmers that will get flagged as errors. However, assembly tools will generally do not do a good job of assembling lowly expressed transcripts anyway.
Why are transcriptome assemblers different from genome assemblers?
Some of the factors that lead to this fragmentation are sequencing errors, polymorphism, sequence repeats, and for more lowly expressed transcripts, stochasticicty of read depth that leads to gaps in coverage. Transcriptome assemblers, unlike genome assemblers, must handle the wide range of depth of coverage due to gene expression variation.
What’s the best way to trim a transcriptome?
First, install TrimGalore! into your home directory where you keep software. Then, make and output directory for your trimming results, create symlinks of the Rcorrector-corrected fastq files (with unfixable reads removed), and submit the trimming job with an sbatch script:
When to trim bases in Illumina transcriptome assembly?
The Illumina demultiplexing pipeline may incompletely remove adapter sequences, and when the insert sizes for a give read pair lead to overlaps between the sequenced bases, sequencing for one read can extend into the adapter of the other. These bases, as well as low quality bases should be trimmed prior to running Trinity.
Why are erroneous k mers bad for transcriptome assembly?
Because the current state of the art transcriptome assemblers use a DeBruijn Graph approach that constructs graphs from kmers, erroneous k-mers can adversely impact assemblies. Because rare k-mers are likely due to sequencing errors, correcting reads such that rare k-mers are corrected to a more frequently occurring can improve assemblies.